请问泊松回归结果中自动报告的alpha代表什么(面板数据),另外,对于泊松回归是选择混合、固定还是随机好怎么判断呢?
随机效应结果给出了alpha ,混合和固定没有给alpha,固定还没有给出常数项的值为什么?
混合
Poisson regression Number of obs = 726
LR chi2(6) = 22.36
Prob > chi2 = 0.0010
Log likelihood = -54.114426 Pseudo R2 = 0.1712
--------------------------------------------------------------------------------
stable | Coef. Std. Err. z P>|z| [95% Conf. Interval]
---------------+----------------------------------------------------------------
lndensity | .5195485 .2681502 1.94 0.053 -.0060162 1.045113
lndensity_2 | .0063155 .0252723 0.25 0.803 -.0432174 .0558484
lngdp | .7109431 .2985539 2.38 0.017 .1257882 1.296098
minorate | .0005779 .0001725 3.35 0.001 .0002399 .000916
lnbtdis | -.1560401 .2641374 -0.59 0.555 -.6737399 .3616597
lnbtdis_stroad | .00041 .0011578 0.35 0.723 -.0018592 .0026792
_cons | -11.79253 3.62797 -3.25 0.001 -18.90322 -4.681834
--------------------------------------------------------------------------------
随机
Random-effects Poisson regression Number of obs = 726
Group variable: code Number of groups = 57
Random effects u_i ~ Gamma Obs per group: min = 1
avg = 12.7
max = 23
Wald chi2(6) = 14.64
Log likelihood = -52.656942 Prob > chi2 = 0.0232
--------------------------------------------------------------------------------
stable | Coef. Std. Err. z P>|z| [95% Conf. Interval]
---------------+----------------------------------------------------------------
lndensity | .5797548 .3577717 1.62 0.105 -.121465 1.280974
lndensity_2 | .0102736 .0322388 0.32 0.750 -.0529134 .0734605
lngdp | .7648429 .356205 2.15 0.032 .066694 1.462992
minorate | .0005606 .0002015 2.78 0.005 .0001657 .0009555
lnbtdis | -.130738 .2633527 -0.50 0.620 -.6468998 .3854237
lnbtdis_stroad | .0003005 .0012208 0.25 0.806 -.0020922 .0026933
_cons | -12.31622 4.179819 -2.95 0.003 -20.50851 -4.123921
---------------+----------------------------------------------------------------
/lnalpha | .3741837 1.014695 -1.614581 2.362949
---------------+----------------------------------------------------------------
alpha | 1.453804 1.475167 .198974 10.62223
--------------------------------------------------------------------------------
Likelihood-ratio test of alpha=0: chibar2(01) = 2.91 Prob>=chibar2 = 0.044
固定
xtpoisson stable lndensity lngdp minorate lnbtdis lnbtdis_stroad,fe
note: 2 groups (2 obs) dropped because of only one obs per group
note: 47 groups (614 obs) dropped because of all zero outcomes
Iteration 0: log likelihood = -29.507269
Iteration 1: log likelihood = -25.458382
Iteration 2: log likelihood = -25.317224
Iteration 3: log likelihood = -25.316112
Iteration 4: log likelihood = -25.316112
Conditional fixed-effects Poisson regression Number of obs = 110
Group variable: code Number of groups = 8
Obs per group: min = 3
avg = 13.8
max = 23
Wald chi2(5) = 7.07
Log likelihood = -25.316112 Prob > chi2 = 0.2154
--------------------------------------------------------------------------------
stable | Coef. Std. Err. z P>|z| [95% Conf. Interval]
---------------+----------------------------------------------------------------
lndensity | -.035994 .5945232 -0.06 0.952 -1.201238 1.12925
lngdp | 1.388761 .6927409 2.00 0.045 .0310135 2.746508
minorate | -.0001829 .0006255 -0.29 0.770 -.0014088 .001043
lnbtdis | .023634 .268513 0.09 0.930 -.5026417 .5499098
lnbtdis_stroad | -.0009434 .001639 -0.58 0.565 -.0041558 .0022689
--------------------------------------------------------------------------------
求大神解惑